Back to glossary

LIME

Local Interpretable Model-agnostic Explanations; a technique that explains the predictions of any machine learning classifier by approximating it locally with an interpretable model.

AbbreviationLIME
Categoryai

Overview

LIME helps us understand 'black box' models by looking at what happens to the prediction when we slightly change the input data.

How it Works

  1. Pick a specific data point you want to explain.
  2. Create a new dataset of 'perturbed' samples (slightly modified versions of that point).
  3. Get predictions for these samples from the complex model.
  4. Train a simple, interpretable model (like a linear regression) on this new dataset.
  5. The simple model's weights provide an explanation of the complex model's behavior around that specific point.

Comparison with SHAP

LIME is generally faster but can be less stable and mathematically rigorous than SHAP.

Related Terms

SHAPSaliency Map